[PATCH] D82248: AMDGPU: Don't ignore carry out user when expanding add_co_pseudo

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: rampitec, alex-t.
Herald added subscribers: kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, jvesely, kzhuravl.
Herald added a project: LLVM.

This was resulting in a missing vreg def in the use select
instruction.

      

copyPhysReg changes lifted out of
4067de569f119a81419fbf2e79d5f3307dfdda5b <https://reviews.llvm.org/rG4067de569f119a81419fbf2e79d5f3307dfdda5b>, which was reverted earlier
today.

      

The output of the pseudo doesn't make sense, since it really shouldn't
have the vreg output in the first place, and instead an implicit scc
def to match the real scalar behavior.

      

We could have easier to understand tests if we selected scalar
versions of the [us]{add|sub}.with.overflow intrinsics.

      

This does still end up producing vector code in the end, since it gets
moved later.


https://reviews.llvm.org/D82248

Files:
  llvm/lib/Target/AMDGPU/SIISelLowering.cpp
  llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
  llvm/test/CodeGen/AMDGPU/expand-scalar-carry-out-select-user.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D82248.272218.patch
Type: text/x-patch
Size: 6628 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200620/fe29d8b6/attachment-0001.bin>